posted on Sat, Jan 18 '20 under tag: events

IndiaOS is an open-source conference organized by ERPNext and Zerodha. This is my report from attending it (as much as I recall from memory)1

I was excited to attend IndiaOS from the first day I saw the website. After foss.in there has not really been a conference of the kind and IndiaOS could one day become that. (Not that I have ever attended FOSS.in)

I reached the venue at 08:25 and got the seat next to a power socket. I was so early that people from either organizing organizations thought I was a member of the other organization. There were folks from Zerodha, and folks from Frappe (who build ERPNext). Labeeb, one of the builders of ChaliMachine who is at Zerodha now, welcomed me. Joice joined soon and told me about Anand Chitipothu that Anand has worked with Aaron Swartz on web.py. I immediately tracked him down over tea (yes, this event had welcome tea - one of the many things they did perfectly) and talked to him about his latest work. Instead, he told me about his older work and turns out he had worked at Strand right from the beginning and he knew exactly what I was talking about just as I began to explain what I do at Metastring. World is a small place.

Soon everyone was coming in and the very beautiful Vasanthi Hall of Balan Farm was full. There were some 18 talks to be made and therefore everything started on time. In fact, the time keeping was perfect till past lunch break.

First Rushabh Mehta, the founder of Frappe talked about the idea of IndiaOS. How it was planned to be a small get-together but the response was incredibly overwhelming and they had to look for bigger venue. He talked about building Frappe. He talked about FOSS. Hrishi and Hiran came in during about that time. And soon many others from SMC were making the last two rows our temporary headquarters.

Then, Dr Kailash Nadh - doctorate doctor, not the other kind - who is the CTO of Zerodha gave an inspiring speech to all hackers. He spoke about how Zerodha is a validation of the hacker culture. About how they took decisions like adopting flutter when it was in alpha stage (for Kite and others, mind you!). Incredible story it is. The founder who is incredibly supportive of the tech team. And the tech team who is not afraid of technology. Many big lessons. There is open source everywhere. Copyright of projects where a single dev makes most contributions is assigned to that dev (and not the compnay!). It is even namespaced under the respective dev. Self-host everything. Not be afraid to build things (like list manager2) that are not in direct line of business. I asked the question “How do you hire such good people?” and Nadh said that they don’t. They hire hackers. (They find out who a hacker is by looking at what they have built) And they grow together. Many friends in audience told stories of how Zerodha tech was a place where real hackers could join without having to answer questions like “What is Big-O notation?”. And did you know, Zerodha is bootstrapped?

Pranav Raj talked about open-sourcing a failed multi-channel customer support software.

Sanket Sarang talked about BlobCityDB which sounds like it has some amazing capabilities in indexing multiple datatypes (like excel files, for example) and that too intelligently.

Shahidh K Muhammed replaced Tanmai Gopal and talked about Hasura which is another of those promising projects. Hasura allows you to build a GraphQL API out of your traditional RDBMS with least effort. And just before I gave up on Mozilla I was coordinating community partnership for Hasura’s GraphQL Asia 2020 conference (the tickets of which are still available).

The lunch was super. There was butter rotti and palak paneer and pulav and many other dishes. I ate twice what I usually eat. And there was a lot of talking. I asked Pranav if Chatwoot really supported WhatsApp and he told how they work with a partner from Indonesia who has access to WhatsApp’s closed business API and use it to get access for bigger clients. Apar and Zainab were talking about tech policy. There was a circle of people who were sharing the most exaggerated claims of “Artificial Intelligence” their PR team makes over Math.random(). I met Sachin Titus, Ranjit Ramakrishnan, Deepthi JS, Nemo, Biswas Babu and many others whose names don’t all come to me at once.

SMC took a group photo. (I’m embedding the Telegram post which has the photo here. But you may not be able to see it, depending on how screwed up the censorship is in your place)

The first talk post lunch was a really important one about free hardware and freeing the entire stack by Abhas of deeproot GNU/Linux. Unfortunately I couldn’t concentrate on this one. The main point was about being careful about the stack we run on, I think. There is a summary on the proposal page.

Mohammed Niyas from gramener made an interesting talk about mapping India. There were some interesting bits about how there are different geographical hierarchies in our country, but not a lot of official shape files for these. He did mention datameet’s work on this. He demoed a time-series of temperature map of India and showed some interesting patterns. And then he talked about a library called reshaper which seems to be a wrapper to geopandas and scipy.spatial that makes it easy to create various splits of map polygons. Going to github made me find yet another Anand who seems interesting. A claim was also made on using all this to gather health insights. I asked where they get data from and the response was “NHFS” (sic). All that I will say is that that’s the best we have.

There were two quick talks about Frappe as a platform and Frappe Books built using FrappeJS.

Then there was the most unwelcome talk of the entire day by Anadi Mishra of MOSIP. The first part of it was exaggerated claims of the benefits of Aadhaar which was discontinued after audience members called out the bull-shit and asked him to focus on the tech. What was supposed to be a lightning talk of 15 mins was already becoming too long. And then the speaker actually said “Okay, coming back to MOSIP” and started talking about this openwashing product. He quickly wrapped with an invitation for all of us to work on it. Someone else asked him about the design flaw in making primary key known to users. I stood up and said we won’t contribute to an “open-source” project for “other countries” when the entire aadhaar system that applies to Indians is not open source. The MOSIP thing deserved no space in IndiaOS as there is nothing that most Indians will gain from MOSIP whereas it is also being used to make India’s Aadhaar look good. Good that the conversation happened. In fact, the controversy was so strong on this one that nobody really bothered to listen to the next talk about an e-commerce store that used Frappe lots of open source3.

Then there was this heart-warming talk by Biswas about the website that was quickly set up by a group of students under IEEE in response to the Kerala floods of 2018 and how the gravity of the situation put on them the responsibility, the thrill, and the opportunity to help rescue Kerala. The deep silence during critical parts of the talk and the unending applause after the talk is enough proof that even if Biswas gives this talk 10 more times, it will still be worth listening to.

Next, our own Anish Sheela talked about the handwriting recognition project that SMC has been building and invited more polyglot hands to contribute.

Somewhere in between we had the evening tea.

Kiran talked about Hasgeek’s journey and made the proposition that if your business relies on your code being closed-source you probably have a shitty business. He shared an old blog post that was relevant. I talked with him later to figure out if he meant that there is absolutely no business where technology is valuable. He agreed that there is a segment where technology is the core IP and there you would want things like software patent to protect it. But for a large number of businesses, technology is something that can easily be copied by others, yet beating the business is not as easy and requires winning over the community and everything else. We considered the case of a small startup building a successful project/product that large companies (with all their resources) are not able to copy. This is because it is difficult for big companies to be nimble and agile. But this is also because big companies have different kinds of metrics. For example, Google may shut down a project that won’t have a million users. But smaller startups may operate on different levels of satisfaction. And there is also the ability to maintain someone else’s project as a fork. Lots of thoughts there. “Talk is cheap, show me the code” said Linus. “Code is cheap, show me the talk” says Kiran.

Mujeeb of Alpha Fork gave the inspiring story of converting an entire newspaper group of windows users into a custom Kubuntu version that runs free software for everything from typesetting to printing.

Rishabh Nambiar gave a lightning talk about discourse. Turns out, discourse has an entirely remote team like gitlab. It is also nice to know that there are bigger advocates of discourse than me.

Anivar had come by then and gave the talk about the trends in software freedom licensing. He mostly focussed on the bad trends like open washing, license changes, etc (without calling them bad) and made the point that copyright and copyleft are but tools and that tools can be wielded in the right way and the wrong way. He also talked about how licensing and compliance should be discussed with engineers first, and lawyers last.

And that is the same point Apar made. Apar talked about IFF’s work and also about how hard it is to make legal changes and more importantly, why engineers should engage.

At the very end, there was the open discussion. Hrishi rightly made the point about lack of gender diversity. The organizers first tried to defend by saying they tried, but later agreed to do better next time. I chipped my tooth in my eagerness to grab the mic to state that “women not filling the call for proposals” is an invalid response. But Kiran put that point in better, talking about systematic ways women are excluded and the need for doing better.

Indeed everyone agreed that the conference felt like it was “professionally” organized. Everyone wants this to happen again. And I’m sure many went back motivated to write more “free” lines of code.

  1. There are probably many many errors in this post. (I’ve already corrected two after publishing. Please let me know if you find one.) 

  2. If you are interested in subscribing to my newsletter, send me an email 

  3. https://twitter.com/rushabh_mehta/status/1219110240056004608 

Like what you are reading? Subscribe (by RSS, email, twitter, or telegram)!